Liquefaction Susceptibility Mapping in Boston,

CHARLES M. BRANKMAN1 William Lettis & Associates, Inc., 1777 Botelho Drive, Suite 262, Walnut Creek, CA 94596

LAURIE G. BAISE Department of Civil and Environmental Engineering, Tufts University, 113 Anderson Hall, Medford, MA 02155

Key Terms: Liquefaction, Seismic Hazards, Engineer- area, and it will assist in characterizing seismic ing Geology, Geotechnical, Surficial Geology hazards, mitigating risks, and providing information for urban planning and emergency response.

ABSTRACT INTRODUCTION The Boston, Massachusetts, metropolitan area has Boston, Massachusetts, is located in a region of experienced several historic of about moderate historic seismicity, where several historical magnitude 6.0. A compilation of surficial geologic events of about M6.0 have occurred (e.g., 1727, 1755). maps of the Boston, Massachusetts, metropolitan area The possibility therefore exists for the generation of and geotechnical analyses of Quaternary sedimentary -induced liquefaction of near-surface sedi- deposits using nearly 3,000 geotechnical borehole logs ments in the Boston area. In this paper, we present reveal varying levels of susceptibility of these units to results of a study to assess the liquefaction suscepti- earthquake-induced liquefaction, given the generally bility of natural sediments and areas of artificial fill in accepted design earthquake for the region (M6.0 with the Boston metropolitan area, with the aim of 0.12g Peak Ground Acceleration (PGA)). The majority characterizing liquefaction hazard and providing of the boreholes are located within the extensive information to local communities for improved downtown artificial fill units, but they also allow planning and mitigation strategies. The primary goal characterization of the natural deposits outside the of the study was to develop liquefaction susceptibility downtown area. The geotechnical data were comple- maps by combining surficial geologic mapping with mented with surficial geologic mapping, combining subsurface borehole data. To develop these maps, published and unpublished geologic maps, aerial existing surficial geologic maps at various scales were photographic interpretation, and soil stratigraphy data augmented with field reconnaissance mapping to from an additional 12,000 geotechnical boring logs. provide a base for assessing the properties of the Susceptibility maps were developed based on liquefac- geologic units. An extensive digital borehole database tion-triggering threshold ground motions, which were was compiled to provide data on the subsurface determined using the borehole data. We find that much properties; it is composed of nearly 3,000 borings, and of the non-engineered artificial fill that underlies the it focuses on the artificial fill units in downtown downtown Boston area is, when saturated, highly Boston but also provides coverage of the other susceptible to liquefaction during seismic loading. geologic units. The subsurface properties, including Holocene alluvial and marsh deposits in the region soil type, standard penetration test blow counts, and are also moderately to highly susceptible to liquefac- estimated fines content, were used to determine tion. Much of the outlying area is underlain by liquefaction susceptibility of each individual sample Pleistocene and Quaternary glacial and glaciofluvial in the database. The liquefaction susceptibility deposits, which have low to moderate susceptibility to mapping used the results of both the surficial geologic liquefaction. This study provides data needed to mapping and the subsurface sample liquefaction effectively manage liquefaction hazards in the Boston susceptibility analysis. The study area encompasses eight 1:24,000-scale 1Present address: Department of Earth and Planetary Sciences, Harvard University, 20 Oxford Street, Cambridge, MA 02138; phone: (7.5 minute) quadrangles in the metropolitan Boston 617-495-0367; fax: 617-495-7660; email: [email protected]. region, and it includes the downtown Boston area and

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 1 Brankman and Baise

Figure 1. Quadrangle outline map of the greater Boston region showing community boundaries. surrounding communities (Figure 1). Much of the southern Quebec had a magnitude of about 7.0 and region is underlain by Pleistocene and Quaternary caused ground shaking intensity in Boston of at least glacial till and glaciofluvial deposits, as well as large V–VI, resulting in damage to several chimneys in the areas of marsh deposits and extensive regions of non- Boston area (Crosby, 1923; Ebel, 1996). The 1727 engineered artificial fill. Based on their composition Newbury earthquake occurred approximately 56 km and conditions of geologic deposition, glaciofluvial northeast of Boston, with an estimated moment deposits, marsh deposits, and especially the non- magnitude (Mw) of 5.6, a reported local MMI of engineered artificial fill are potentially susceptible to VI–VII, and a MMI for Boston and northern suburbs liquefaction during large earthquakes. of V–VI (Ebel, 2000). Reports in Newbury at the time of the earthquake describe sand boils, which indicate liquefaction (Plant, 1742; Ebel, 2000). These occur- BACKGROUND rences of liquefaction have been confirmed by Seismic History paleoseismic studies, which found sand dikes and sills in glaciomarine sediments in two locations The Boston region has experienced several historic corresponding to the liquefaction during the 1727 earthquakes that have caused ground motions signif- earthquake and one prehistoric event (Tuttle and icant enough to trigger liquefaction in susceptible Seeber, 1991; Tuttle et al., 2000). sediments. In 1638, an earthquake thought to have The 1755 Cape Ann earthquake was the largest been located in central struck with earthquake to have affected Boston in historic times, a magnitude (MbLg) of about 6.5; Ebel (1996, 1999) and it caused damage throughout eastern Massachu- estimated that the event produced modified Mercalli setts and was felt along the eastern seaboard of North intensity (MMI) of V–VII in Boston. An earthquake America from Nova Scotia to (Ebel, in 1663 located within the Charlevoix seismic zone in 2006). The earthquake was located approximately

2 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 Boston Liquefaction Susceptibility Maps

40 km ENE of Cape Ann, Massachusetts, and it had interpretations of stratigraphy derived from over aMw of about 5.9 (Ebel, 2006). The earthquake 5,000 boring logs. The hazard classification for the caused extensive damage in Boston, destroying at Victoria maps was based on an interpretation of the least 1,500 and as many as 5,000 chimneys (Whitman, stratigraphy represented in the boring logs and 2002), and it reportedly affected water levels in wells a detailed analysis of 31 sites. The detailed analysis as far away as central and western Connecticut consisted of a combination of a probabilistic pre- (Thorson, 2001). Crosby (1923) estimated that the diction of liquefaction using the Seed and Idriss 1755 earthquake caused a MMI of IX in Boston, (1971) simplified approach and a probability of while Ebel (2006) estimated a MMI of VII. Estimates liquefaction severity index, which depends on depth of ground motions in Boston range from 0.08 to 0.12g and thickness of the liquefiable materials (Monahan (Ebel, 2006). Written accounts of damage caused by et al., 1998, 2000). the 1755 earthquake in Scituate, about 30 km Recently, the California Geological Survey (CGS) southeast of Boston, reported liquefaction sand boils; has produced seismic hazard zone maps that delineate these features were investigated using paleoseismic areas that are likely to contain liquefiable sediments and geophysical techniques, but the studies were not in seismically active areas of the state. The CGS conclusive (Tuttle et al., 2000). zonation is based on susceptibility evaluations that use geologic criteria and borehole analyses similar to Liquefaction Hazard Mapping the method used in this study. To date, CGS has compiled Quaternary geology for 113 U.S. Geological Regional liquefaction hazard mapping projects Survey (USGS) 7.5-minute quadrangles and has have predominantly relied on criteria that relate collected and analyzed over 16,000 borehole logs Quaternary surficial deposits to liquefaction suscep- from the greater Los Angeles and San Francisco Bay tibility, taking into account factors such as deposi- area (California Geological Survey, 2007). tional environment, dominant grain size, and relative age (Youd and Perkins, 1978). This methodology commonly leads to the identification of large regions METHODOLOGY of susceptible material. Youd and Perkins (1987) We applied regional-scale liquefaction mapping discussed how the resulting maps show geologic units criteria based on surficial geology and analysis of that likely contain liquefiable sediments but do not geotechnical data to prepare liquefaction hazard identify the precise location of the liquefiable maps for the greater Boston metropolitan region. sediments within the geologic unit. Therefore, it is The mapping criteria consisted of three hazard classes possible that within a susceptible unit only a small (low, moderate, and high), which refer to varying discrete area or areas will actually liquefy during extents of expected liquefaction. Our intent was to a given earthquake. provide classes of hazard based on both geologic and Recent liquefaction mapping projects have typical- geotechnical criteria that account for the variability of ly included the concurrent collection of subsurface geologic materials as well as the distribution of data to provide more quantitative susceptibility liquefiable materials within an individual geologic estimates. The subsurface data may include standard unit. penetration test N-values, cone penetrometer (CPT) data, shear-wave velocity (Vs), soil descriptions (including grain-size distributions), stratigraphy, and Surficial Geologic Maps groundwater measurements. Hitchcock et al. (1999) conducted extensive investigations of liquefaction Surficial geologic maps of eight 1:24,000-scale hazards and produced detailed susceptibility maps quadrangles (Figure 1) were compiled from existing in Simi Valley, California, using surficial geologic published geologic maps, where available, and these mapping and analysis of over 1,000 boring logs from were augmented with reconnaissance field mapping a variety of Quaternary deposits. Hitchcock and throughout the study area. High-quality, large-scale, Helley (2000) collected over 1,600 boring logs for 12 published maps were available for the Norwood 7.5-minute quadrangles in the Santa Clara Valley, (Chute, 1966) and Blue Hills (Chute, 1965) quad- California. The boring logs were used to help rangles, and for portions of the Boston North and delineate the top of the Pleistocene deposits, estimate Lexington quadrangles (Chute, 1959). Smaller-scale the thickness of Holocene sediments, and determine maps of the entire study area were available (e.g., the thicknesses and time of placement of artificial fills. Kaye, 1978; Thompson et al., 1991; and Woodhouse Monahan et al. (2000) produced relative liquefaction et al., 1991) and provided a first-order base map for hazard maps for Victoria, British Columbia, from use in field checking.

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 3 Brankman and Baise

In the mapping, we faced two primary challenges. quadrangles (e.g., the Reading quadrangle; Oldale, First, the area is extensively developed; exposure is 1962) allowed us to check geologic contacts along the typically less than one to two percent, and there has quadrangle boundaries. Finally, and importantly, we been large modification of the land surface through- were able to confirm the map units and refine unit out the study area. Extensive grading for construc- contacts using data from the borehole database and tion, draining and filling of wetlands and marshes, the larger database of borings from the Massachusetts channeling and diversion of streams and rivers, and Water Resources Authority. modification of river banks have occurred over the past three centuries. These cultural processes often obscure the nature of the underlying deposits, and Geotechnical Borehole Database they occur not only in the densely populated downtown Boston and surrounding urban areas, but Data for this project were acquired from several also in the outer suburban regions. This difficulty sources. An electronic collection of data (1,905 directly affected the level of detail that could be borings) was acquired from the Central/Artery attained in subdividing units during the surficial Tunnel project in Boston through the Massachusetts mapping. Second, previous workers mapping the Water Resources Authority (MWRA). This database region over the past century have adopted a variety was modified from the original to fit into a standard of classification schemes for the surficial geologic format. Geologic descriptions varied considerably units. This can be attributed to both the development and were therefore simplified to be more consistent of the science of glacial and Quaternary geology over throughout the region, though sample information the past century and also the wide variety of scales of was not altered. In addition, electronic scanned mapping and the various locales that were the focus images of 12,782 boring logs and their location of the mapping projects. coordinates were acquired from the MWRA. Due to We addressed these issues by using generalized the large number of these data, we selected borings geologic units based on those defined by Chute (1965, from this set to fill coverage gaps in the other boring 1966). We divided surficial units into six general units, databases. Data from 119 of these boring logs were including glacial drumlins (glacial till), glacial ground hand-entered into the database. Additional MWRA and end moraines (glacial till), glaciofluvial deposits logs were examined as needed for the surficial (glacial outwash plains, eskers, kames, and kame geologic mapping and assessment of map unit fields), marsh deposits, beach deposits, and historic boundaries. The Boston Society of Civil Engineers artificial fill. These units, while general, group (BSCE, 1969) collection of borings was also used as deposits based on common depositional processes, a data source, and 314 borings from the BSCE composition, and age, and they are present through- collection were hand-entered into the database. out the study area. In addition, these units form Finally, for the Cambridge area, 715 borings were relatively distinct geomorphological terrains and can collected in and near the Cambridge fill unit along the be identified with confidence on the basis of their northern shore of the Charles River. The resulting surface expression. This allowed us to map geologic geotechnical database from all of the aforementioned units even with the lack of exposures described sources includes 2,963 borings. previously. Admittedly, there is variability of geologic Data from geotechnical borings were entered into properties within each unit, and in some cases, our an electronic database in order to facilitate relational morphology-based mapping may have passed over database management and allow for the flexibility of some of the details of the contacts between adjacent data input. The database includes both general and map units. However, given the challenges imposed by geologic information gathered from subsurface ex- the issues as described here, we feel that these unit plorations, such as project and drilling information, designations do not introduce substantial error into date and depth of boring, ground surface elevation, the mapping and provide a good base map for the depth to groundwater, depths and descriptions of liquefaction analyses. stratigraphic units and samples, standard penetration Validation and confirmation of our mapping were test (SPT) N-values, and x-y coordinate values. The accomplished by performing reconnaissance mapping soil samples were characterized by soil type (i.e., sand, of portions of quadrangles with published surficial silt, silty sand, clay, etc.) and a detailed sample geologic maps prior to examination of those pub- description. When available on the original boring lished maps and then comparing the interpretations log, stratigraphic unit was also associated with between the maps. In all cases, our reconnaissance individual soil samples. The stratigraphy was charac- mapping provided good agreement with the published terized by geologic unit and depth to top and bottom maps. In addition, published maps from adjacent of each unit. In some cases, the stratigraphic unit was

4 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 Boston Liquefaction Susceptibility Maps modified slightly from the original boring log in order characterization on two scales: regionally, based on to conform to a uniform naming convention. surficial geologic unit, or locally, based on SPT data. We used a design earthquake of Mw 5 6.0 for the scaling factors used in the trigger-level calculations to Quantitative Analysis assess the liquefaction susceptibility, based on the historic earthquake record and, specifically, the Liquefaction susceptibility refers to the relative magnitude of the 1755 Cape Ann event (see previous). resistance of soils to loss of strength due to an In addition, we selected a PGA of 0.12g as our design increase in pore-water pressure caused by ground ground motion for determination of liquefaction shaking. The degree of resistance is governed triggering. Using the 2002 and 2007 (proposed) primarily by the soil’s physical properties such as USGS Probabilistic Seismic Hazard Maps, values grain size, density, and saturation. Zones correspond- for two percent in 50 years peak ground acceleration ing to areas of very low to very high susceptibility can for Boston match the chosen design value of 0.12g be defined based on a liquefaction-triggering thresh- (USGS, 2007). For Boston, the Massachusetts Build- old analysis using standard penetration test (SPT) ing Code mandates a peak ground acceleration of data in areas with borehole data, and with criteria 0.12g, which is consistent with the standard of based on the deposit’s age, texture, and groundwater practice. condition in areas lacking borehole data. In regions where subsurface data were not avail- Where borehole data were available, liquefaction able, we assessed liquefaction susceptibility using susceptibility was quantified according to the adjust- geologic criteria as originally defined by Youd and ed SPT blow count (N1)60 values. This quantitative Perkins (1978). These criteria are based on the evaluation of liquefaction susceptibility was based on physical characteristics of a given geologic unit that the Seed-Idriss simplified procedure, which was impact the susceptibility of that unit to coseismic reviewed and updated in a workshop report summa- liquefaction, including the depositional environment, rized by Youd et al. (2001). This procedure calculates age, lithologic composition, grain-size distribution, soil resistance to liquefaction, expressed in terms of density, and degree of saturation. While inherently cyclic resistance ratio (CRR), based on SPT data, qualitative, classifications using these criteria have groundwater level, soil density, percent fines, and been verified through the response of similar geologic sample depth. The groundwater levels in Boston are units to ground motions during recent large earth- highly locally variable as a result of sewer systems, quakes, and it is appropriate for regions without dewatering projects, and seasonal variations. We used available subsurface data. groundwater data where noted in the boring logs; otherwise, we used a conservative groundwater level at the ground surface. CRR values were compared to Liquefaction Hazard Mapping Methodology cyclic shear stresses generated by the estimated In moving from a local, sample-scale assessment of ground motions, expressed in terms of cyclic stress liquefaction susceptibility to a regional-scale suscep- ratio (CSR). Appropriate correction factors for SPT tibility map, we used a combined approach using both values, and scaling factors for fines content and the geologic criteria described already and statistical earthquake magnitude (see following), were applied analysis of the subsurface data. The statistical as suggested in Youd et al. (2001). For each soil methodology is described more completely in Baise sample in the database, a liquefaction trigger level of et al. (2006), and Table 1 summarizes the liquefaction the peak ground acceleration (PGAtrigger) was calcu- hazard mapping criteria used in this study. lated using the simplified Seed-Idriss approach (Youd In the development of the liquefaction hazard et al., 2001), as described already, which takes into mapping criteria, an investigation of population account depth, saturation, soil type, density, and fines statistics was completed for several liquefiable depos- content. A factor of safety (FS) equal to 1.2, which its in the study area. The accuracy of the character- has been recommended to achieve a 20 percent ization depended primarily on the amount of data probability of liquefaction (Juang et al., 2002), was available. Using liquefaction probabilities instead of used in the calculations. Thus, we calculated the susceptibility categories, Baise et al. (2006) found that PGAtrigger for each soil sample in the database, which the estimated error for the liquefaction probability allowed us to provide individual classifications of over a subsurface unit was directly related to the susceptibility. These susceptibility category values are mean value and the number of samples. Susceptibility distinct from the geologically determined criteria estimates from population percentages based on because they are specific to an individual soil sample PGAtrigger values can be misleading when based on rather than the entire geologic unit. This allows a small number of samples from within a given

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 5 Brankman and Baise

Table 1. Hazard criteria used in this study and regional susceptibility mapping criteria based on geologic characteristics of various map units (after Youd and Perkins, 1978).

Geotechnical Hazard Category Geologic Criteria (Susceptibility) Geologic Units Boring-Based Criteria High hazard Modern to Holocene; saturated; abundant Artificial fill .20 percent of borings with cohesionless, uncompacted sediments Active beach deposits liquefiable samples Moderate hazard Holocene to Pleistocene; saturated; variable Glaciofluvial deposits 5 to 20 percent of borings with amounts of cohesionless, uncompacted Marsh deposits liquefiable samples sediments Low hazard Pleistocene to pre-Pleistocene; non-saturated Glacial till (drumlin and ,5 percent of borings with to saturated; well indurated; cohesive; ground moraine) liquefiable samples limited cohesionless sediments Bedrock geologic unit. Therefore, broad regional estimates of sequences of fluvial sands and overbank silt deposits, liquefaction susceptibility based on limited samples which line the margins of the river channels and are resulted in large levels of estimate uncertainty. If now often present in the subsurface under the artificial a sufficient sample density is not available, the fill units along the banks. characterization should rely more heavily on the Glacial till is mapped as two separate units: glacial surficial geology. drumlins and ground moraines. Both generally lie The liquefaction hazard criteria presented here do directly on the bedrock surface and were deposited not describe expected deformations resulting from below the advancing glaciers or during the melting of liquefaction (i.e., settlements, lateral spreading, etc.), stagnant or receding ice (Chute, 1966). These two which would depend on thickness of susceptible unit, units were differentiated in the mapping on the basis depth to susceptible unit, lateral extent of susceptible of their differing and unique morphologies. Drumlins unit, surface and intra-unit topography, and nearby are present throughout the study area, and they have structures. Rather, the maps produced using these been well described in the literature (e.g., Woodhouse criteria are meant to characterize the spatial extent of et al., 1991). They occur as round to elliptical hills and liquefiable materials. In addition, the liquefaction highlands generally reaching several tens of meters hazard mapping is meant for the regional scale and above the surrounding terrain. Drumlins are often not the site-specific scale. This information can be cored by local bedrock highs. Prominent drumlins used to plan detailed explorations for a site to confirm include several in the Somerville-Medford-Charles- the liquefaction hazard at the site. town areas north of Boston, and throughout the Boston outer harbor, where drumlins form many of the harbor islands. Ground moraines are also SURFICIAL GEOLOGIC MAPPING composed of glacial till but are generally confined Quaternary Geology to the highlands north, west, and south of Boston. These mapped areas of ground moraine also include The Quaternary geology of the Boston area extensive areas of bedrock exposure in some of the (Figure 2) is dominated by sediments deposited during higher elevations; however, since the areas of bedrock and after extensive and repeated Pleistocene glacia- are often discontinuous and occur almost exclusively tions of the area (Kaye, 1982; Barosh et al., 1989; and within the ground moraine unit, we do not break out Woodhouse et al., 1991). Glacial advances deposited individual areas of bedrock exposure on the maps. till as drumlins and ground moraines, while glacial Rather, we note that the ground moraine unit can withdrawal during the late Pleistocene deposited large vary in thickness from several tens of meters to zero, regions of glacial outwash. The outwash and till and bedrock exposures can occur in zones of zero together comprise about 75 percent of the surface in ground moraine thickness. Where present, the ground the study area. During and after glaciation, coastal moraine till ranges in thickness from zero up to processes influenced by the competing effects of approximately 40 meters, while drumlin till can reach crustal isostatic rebound and eustatic sea-level change over 50 meters in thickness (Chute, 1966; Woodhouse resulted in a complex distribution of coastal estuarine et al., 1991). and tidal marsh sediments. Local beach deposits and The till is generally composed of poorly sorted tidal estuary deposits developed along active coastal sand, gravel, and cobbles in a clay matrix, and it is areas and sheltered marshes, respectively. In addition, generally well consolidated and very dense. Large the Charles River and other smaller rivers (e.g., Mystic cobbles and boulders up to 1 m in diameter occur and Neponset Rivers) and streams locally deposited rarely throughout the till, but are often confined to

6 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 Boston Liquefaction Susceptibility Maps

Figure 2. Quaternary geologic map of the Boston metropolitan area, as compiled from published maps and field reconnaissance (see text for references).

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 7 Brankman and Baise the upper 3–4 m (Woodhouse et al., 1991). Silty presence of either buried boulders or fill that was laminations and well-developed internal structure are subsequently buried by placement of sand during often present, in some places highly disrupted and beach reclamation or stabilization. Geologically, the folded by the motion of the glacial ice (Kaye, 1961; deposition of beach deposits results in loose and often Woodhouse et al., 1991). The till ranges in color from saturated sands; therefore, we argue that beach sands brown to yellow to gray. The till was present in 22 should be mapped as highly liquefiable deposits. borings in the database. SPT blow counts in the till We also recognized several stratigraphic units that were variable but generally ranged from about 20 to occur in the subsurface but do not crop out on the refusal. As a result of their geologic characteristics, surface and thus could not be included in the geologic both the drumlin till and the ground moraine till are maps. These units can be laterally extensive; however, not expected to be susceptible to liquefaction. they generally require relatively dense subsurface The glaciofluvial deposits encompass a variety of boring data to map accurately. An example of one deposits formed by the transport of glacially derived of these units is the famous Boston Blue Clay, which materials, either from the glacier front or by sub- underlies much of the Massachusetts Bay area and glacial flow, including outwash, eskers, kettles, kame has been extensively studied in the past because of its fields, and terraces. These deposits are grouped impact on deep foundations of buildings in the together for mapping purposes, and they are referred downtown area. The Blue Clay is a well-bedded to as glaciofluvial deposits. These are composed deposit of clay, silt, and fine sand formed from the primarily of stratified sands and gravels that are rock flour component of glacial outwash (Wood- heterogeneous in three dimensions as well as in both house et al., 1991), and it is not considered to be at density and consolidation. The glaciofluvial deposits, risk of liquefaction. with thicknesses of meters to tens of meters, often overlie ground moraine till, and in several locations Artificial Fill (e.g., Mystic Lakes–Fresh Pond area), they fill buried bedrock valleys up to about 70 m deep (Chute, 1959). The original settlement of Boston was situated on The outwash units range in color from tan to brown and adjacent to Beacon Hill, a drumlin which formed and yellow, and they tend to be loose to dense. The an island at high-tide (Woodhouse, 1989; Seasholes, glaciofluvial units were encountered in 78 borings, 2003). Due to subsequent urbanization, primarily and reported SPT blow counts ranged from five to during the mid 1800s to early 1900s, non-engineered refusal. The presence of large zones of sand and sandy artificial fill was placed on the adjacent low-lying tidal silt in this unit, as well as zones with relatively low marshes, estuaries, and floodplains adjacent to the blow counts, indicates that the glaciofluvial units may Boston Harbor and the Charles River (Figure 3). The be susceptible to liquefaction. fill has recently been documented Modern marsh deposits are common in the study using historical maps and documents (Ty, 1987; area and occur both as salt marshes and estuaries along Seasholes, 2003). Each episode of land reclamation the coastal areas and as freshwater marshes along used specific source material and a different filling streams and rivers further inland. Marsh deposits are method; therefore, it is useful to break up the fill unit generally composed of fine sands, silts, and clays, with into subunits, which can then be characterized in- abundant peat layers. Thicknesses can reach several dividually. Figure 4 presents the 10 fill units delineated meters. These units are generally loose, with SPT blow in this study: Charlestown, Cambridge, Back Bay, counts generally below 10. Marsh sediments were West Cove, Mill Pond, East Cove, South Cove, South encountered in 81 samples from 18 borings. Urbani- Bay, South Boston, and East Boston. Although tidal zation and suburban sprawl have resulted in a large marsh, estuary, and till deposits directly underlie them, amount of filling of these regions over the last 75 years; these regions are mapped as artificial fill. therefore, artificial fill often overlies loose marsh In general, the fill layer consists of loose to very deposits. We consider the marsh deposits to be dense sand, gravelly sand, or sandy gravel intermixed moderately susceptible to liquefaction based on the with varying amounts of silt, clay, cobbles, boulders, presence of discrete layers of saturated, cohesionless and miscellaneous materials such as brick, ash, rubble, sediments between the more organic strata. trash, or other foreign materials (Ty, 1987; Wood- Beach deposits represent the sediments deposited house et al., 1991). The source materials include both by ongoing modern and historic coastal processes. In granular and cohesive material that was obtained from general, these are composed of sand and gravel and nearby quarries and dumped loosely at sites, generally have thicknesses ranging up to several meters. In without sorting or compaction. As a result, properties a small number of borings, extremely high blow of the fill layer are extremely variable, and blow counts counts within the beach deposits indicated the range up to refusal. In addition, much of the fill is

8 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 Boston Liquefaction Susceptibility Maps

Figure 3. Quaternary geologic map of the Boston downtown area. saturated as a result of a relatively shallow (although collected in each of these units; however, the artificial highly variable) groundwater table. fill unit near downtown Boston and Cambridge was The regions mapped as artificial fill are considered the only geologic unit that was densely sampled. The to have the highest liquefaction potential in the Boston geotechnical data were only sparsely available over area (especially surrounding downtown Boston). If it is the remainder of the study region. Although we used saturated and cohesionless, historic (non-engineered) geotechnical data in all of the geologic units to fill is generally considered susceptible to liquefaction evaluate liquefaction potential, the susceptibility because it was loosely placed. Most of the fill mapping of all units other than the artificial fill relied underlying newer buildings in Boston is engineered fill more heavily on the geologic characterization, as rather than the loosely placed historic fill discussed described next. For the artificial fill unit, a statistical here. Engineered fill, when properly placed and classification was also applied. compacted, is usually dense and not susceptible to In order to use the geotechnical data to ascertain liquefaction under the modest seismic loading expected liquefaction susceptibility, all samples in each of the six in Boston. The historic fill likely remains beneath surficial geology categories were queried from the many historic buildings and roadways. database. Samples from all units in the boring were included in the analyses. The resulting collections of samples included all soil samples taken within the SUBSURFACE CHARACTERIZATION geographic confines of that surficial geologic unit. Liquefaction Susceptibility Susceptibility analyses were run for these samples using the methodology described previously, and The surficial geology maps (Figures 2 and 3) show aPGAtrigger of 0.12g was used as the threshold ground the six geologic units: artificial fill, marsh deposits, motion for determining failure of each sample. The glaciofluvial deposits, drumlin till, ground moraine data and results for each geologic unit are summarized till, and beach deposits. Geotechnical data were in Table 2, and the distributions of susceptibility

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 9 Brankman and Baise

Figure 4. Geographic subunits of artificial fill in the central Boston area. categories for the three geologic units with susceptible taken in the ground moraine till (45 samples from 17 material are shown in Figure 5. As expected, the borings). The ground moraine till is generally a thin artificial fill is the most susceptible unit. Marsh glacial till deposit over bedrock. These deposits are deposits also exhibit a relatively high level of suscep- expected to be dense to very dense. The samples in the tibility. The distribution of liquefiable samples in the ground moraine deposit confirmed the expectation of glaciofluvial deposits exhibits moderate susceptibility. dense soils; therefore, ground moraines till was Very few samples were taken in the drumlin deposit mapped as low hazard. (only 13 samples from six borings across the study Based on depositional environment and field area). Based on the depositional conditions of the evidence, the glaciofluvial deposits are composed glacial till that makes up the drumlins, as well as from primarily of interbedded sand and gravel layers with field observations of exposures of the till, the material some silt and cobble interbeds of variable density. is expected to be very dense and not susceptible to The results from the geotechnical data are variable: 9 liquefaction. Samples from the drumlin deposit out of 79 borings, or 12 percent, have liquefiable confirm this expectation (zero percent liquefiable). material given the design earthquake. Some borings Therefore, all drumlin till deposits were categorized as contain only non-liquefiable samples, while others low susceptibility. Similarly, very few samples were have several samples that would liquefy for a larger

Table 2. Distribution of all analyzed samples by geologic unit characterized as susceptible to liquefaction, given a peak ground acceleration of 0.12g.

Number of Number Percent of Samples Susceptible to the Percent of Borings with at Least One Sample Samples of Borings Design Earthquake (PGA 5 0.12g) Susceptible to the Design Earthquake Artificial fill 9,898 1,727 7.6 29 Marsh deposits 81 18 7.4 22 Beach deposits*** 29 8 0 0 Glaciofluvial deposits 347 79 3.2 12 Drumlin*** 13 6 0 0 Ground moraine*** 45 17 0 0

***Indicates small sample.

10 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 Boston Liquefaction Susceptibility Maps

Figure 5. Histograms showing distribution of susceptible (L) and non-susceptible (NL) samples in liquefaction susceptibility categories for the three mapped geologic units with susceptible samples, for M 5 6.0 and PGAtrigger 5 0.12g. earthquake than the design earthquake (PGA . taken in an actual beach deposit. Most were taken 0.12g). The glaciofluvial deposits were mapped as within historic artificial fill placed along the seashore, moderate hazard; however, if the design earthquake which has been subsequently overlain by beach sand, was altered, the susceptibility could possibly increase. either placed during beach restoration or deposited The liquefiable samples within the glaciofluvial units naturally, and borings have encountered miscella- are isolated, and, therefore, we do not expect large, neous dense materials. Based on geologic criteria, continuous zones of liquefiable materials. natural Holocene beach deposits are expected to be The marsh deposits vary from silty to sandy soils, loose, saturated sandy deposits and are highly with abundant organic layers. Most of the soils in the susceptible to liquefaction. The beach deposits were marsh deposits are loose and saturated. The silty and therefore mapped as high hazard. organic-rich soils are not generally liquefiable; how- Overall, 29 percent of borings in the artificial fill ever, the sandy soils in these deposits tend to be contained at least one liquefiable sample. When we liquefiable in the design earthquake. In the marsh subdivided the fill into the individual subregions as deposits, 22 percent of the borings contain samples summarized in Table 3, the susceptibility was spatial- that are susceptible to liquefaction in the design ly variable depending on the fill and construction earthquake (Table 2). This is similar to the suscepti- history of the area. In several of the fill regions, there bility of the artificial fill (see following). The marsh were large continuous zones of liquefiable materials deposits were therefore mapped as high hazard. (especially West Cove, Mill Pond, Cambridge, Charles- The few samples in the beach deposit were not town, and Back Bay). A detail of Back Bay, with the representative of the sandy soils expected in a Holo- liquefaction categories from the geotechnical borings, is cene beach deposit. After close examination of the 29 shown in Figure 6. On the other hand, South Boston, samples in the beach deposit, none of the samples was South Cove, South Bay, East Cove, and East Boston

Table 3. Distribution of all analyzed samples for geographically designated artificial fill subunits in central Boston, and percent of samples characterized as susceptible to liquefaction, given a peak ground acceleration of 0.12g.

Number of Percent of Samples Percent of Borings with at Least Geographic Number of Number of Liquefiable Susceptible to the Design One Sample Susceptible to the Subunit Samples Borings Samples Earthquake (PGA 5 0.12g) Design Earthquake Back Bay 152 44 25 16 41 West Cove 104 18 14 13 61 Mill Pond 719 116 84 12 46 East Cove 57 13 3 5 15 South Cove 364 121 32 9 18 South Bay 374 130 20 5 14 South Boston 935 220 56 6 21 East Boston 1,453 315 49 3 14 Charlestown 521 110 48 9 35 Cambridge 4,280 638 423 10 40

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 11 Brankman and Baise

Figure 6. Detail of Boston’s Back Bay area showing results of liquefaction susceptibility analyses within the artificial fill. Shaded portions of pie charts represent proportion of different susceptibilities of samples from each boring, for M 5 6.0 and PGAtrigger 5 0.12g. demonstrated a more moderate level of susceptibility, percent of borings within the artificial fill contain though still higher than any other geologic unit in some susceptible soils. The liquefaction susceptibility Boston. is spatially variable across the artificial fill unit and Figure 7 shows the liquefaction susceptibility map includes many continuous zones of liquefiable mate- for the greater Boston area, while Figure 8 shows the rial. The marsh deposits and glaciofluvial deposits distribution of liquefiable samples in the artificial fill have 22 percent and 12 percent of borings with some around the Boston peninsula. The Cambridge fill susceptible material, respectively. The glaciofluvial region is highly susceptible to liquefaction for the sediments were therefore mapped as moderate sus- design earthquake. It should be noted that some ceptibility, while the marsh sediments were mapped as regions of fill, particularly those underlying modern high susceptibility. highways and developments, have most likely been The regional liquefaction susceptibility map for either removed or adequately compacted during Boston and surrounding communities (Figure 7) construction and designed to be resistant to liquefac- shows that the highly susceptible regions are concen- tion. However, because we lacked quantitative geo- trated around downtown Boston, where most of the technical data from most of these site-specific project historic artificial fill and underlying marsh deposits areas, we mapped them like the other non-engineered are located. Although the artificial fill is mapped as fill. Thus, all of the artificial fill regions were mapped high hazard, the material is highly heterogeneous and as high hazard. varies from very loose to very dense. Complete liquefaction of the entire fill region is not likely to DISCUSSION occur; however, large contiguous zones (possibly underlying entire city blocks) are expected. Baise et In greater Boston, the liquefaction susceptibility of al. (2006) presented a detailed case study of the near-surface deposits, both natural and artificial, susceptibility of the artificial fill along the Cambridge varies widely across the region. The artificial fill waterfront area, which demonstrated the spatial units, marsh deposits, and the beach deposits are variability of susceptibility in the fill. While this mapped as high hazard for liquefaction. The beach emphasizes the potential hazard of coseismic lique- deposit characterization is based solely on the faction to structures built in regions of artificial fill, it surficial geology, since the geotechnical data taken is important to note that the downtown Boston area in the beach deposits were sparse and not represen- has been extensively developed, and it is conventional tative. The artificial fill unit in downtown Boston and for large projects to remove the historic fill before in Cambridge has been densely sampled, and 29 construction, or to utilize deep foundations seated in

12 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 Boston Liquefaction Susceptibility Maps

Figure 7. Map showing liquefaction susceptibility of Quaternary geologic units and artificial fill in the Boston region.

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 13 Brankman and Baise

Figure 8. Detail map showing varying liquefaction susceptibility of fill units in downtown Boston. stable material beneath the fill. Therefore, the hazard our analyses, we conservatively assumed a groundwa- to large modern structures is most likely minimal. ter level equivalent to the highest reported level or However, smaller older structures, as well as surface surface groundwater table if none was reported. and near-surface roadways and utilities (lifelines) are Seasonal variations in rainfall and snow meltwater still likely at risk. can be expected to change groundwater levels; as The glaciofluvial sediments were relatively under- a result, susceptibility of given samples from the sampled with respect to the large surface area that boring database may increase or decrease. they cover. Based on the boring data, this unit is only As mentioned already, the susceptibility criteria slightly to moderately susceptible; 9 of 79 borings presented here do not predict expected or possible have at least one sample that is susceptible in the types of deformation resulting from liquefaction, such design earthquake. Where observed in several in situ as lateral spreading or settlements. These largely exposures throughout the study area, however, the depend on the thickness of the susceptible unit, depth sandy portions of the outwash appear to be loose and to susceptible unit, lateral extent of susceptible unit, unconsolidated, and they often comprise laterally surface and intra-unit topography, and nearby extensive and connected deposits. Given these ob- structures, and as such, they are local, site-specific servations, and assuming that similar conditions exist effects that must be considered individually. The in the subsurface, it is possible that the boring data do susceptibility maps produced in this study are meant not adequately characterize the susceptibility of this for the regional scale and not the site-specific scale. unit, and that significant portions of the glaciofluvial This information can and should be augmented with deposits could be at risk for liquefaction. According- detailed site explorations to confirm the liquefaction ly, we assigned the unit a susceptibility rating of hazard at a specific site. moderate. Additional boring data from throughout A primary shortcoming of this mapping project this unit would help to better constrain the potential was the lack of data in large regions of the study area, response of this unit during coseismic ground particularly in outlying areas. For this study, the most motions. susceptible unit is the artificial fill; therefore, only Groundwater level has a primary control on the a limited effort was made to collect geotechnical data liquefaction susceptibility of a given sediment unit. In over greater Boston. The maps are therefore pre-

14 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 Boston Liquefaction Susceptibility Maps dominantly based on surficial geology, although the expressed or implied, of the U.S. government. We underlying classifications are supported by the thank Rebecca Higgins and Kevin Dawson (Tufts quantitative geotechnical data. University) for their assistance in subsurface data collection, and Christopher Hitchcock (W.L.A.) for technical assistance and review. We also thank Ralph CONCLUSIONS Loyd and two anonymous reviewers for extremely We applied the regional liquefaction mapping thorough and constructive reviews and comments, criteria presented in Table 1 to prepare liquefaction which greatly improved the manuscript. The sub- hazard maps for greater Boston. The mapping criteria surface data collected over the course of these consist of three hazard classes (low, moderate, and investigations are available to interested individuals; high) that refer to varying expected extents of please contact Laurie Baise ([email protected]) liquefaction. The criteria are based on surficial for an electronic copy of the database. geology and geotechnical data. The intention of the mapping criteria was to provide a hazard class that REFERENCES accounted for the variability of geologic materials and the distribution of liquefiable materials within a re- BAISE, L. G.; HIGGINS, R. B.; AND BRANKMAN, C. M., 2006, gional geologic unit. Liquefaction hazard mapping—Statistical and spatial char- We assembled surficial geologic maps for the acterization of susceptible units: Journal Geotechnical Geoen- vironmental Engineering, Vol. 132, No. 2, pp. 705–715. greater Boston area (Figures 2 and 3). The maps BAROSH, P. J.; KAYE, C. A.; AND WOODHOUSE, D., 1989, Geology of were developed from existing high-quality, large- the Boston basin & vicinity: Civil Engineering Practice, scale, published maps, smaller-scale maps of the Vol. 4, No. 1, pp. 39–52. entire study area, as well as field reconnaissance BOSTON SOCIETY OF CIVIL ENGINEERS, 1969, Boring data from mapping using field exposures and geomorphological greater Boston: Journal Boston Society of Civil Engineers, Vol. 56, No. 3–4, pp. 131–293. interpretation. To complement the surficial geologic CALIFORNIA GEOLOGICAL SURVEY, 2007, Seismic Hazards Zonation maps, we assembled an electronic database of geo- Program,available at http://www.conservation.ca.gov/cgs/shzp/ technical data from 2,963 test borings. The geo- CHUTE, N. E., 1959, Glacial Geology of the Mystic Lakes–Fresh technical data included stratigraphy, soil sample Pond Area, Massachusetts: U.S. Geological Survey Bulletin description, soil type, groundwater level, and SPT 1061-F, pp. 187–216. blow count. Although the data are concentrated in CHUTE, N. E., 1965, Surficial Geologic Map of the Blue Hills Quadrangle, Norfolk, Suffolk, and Plymouth Counties, Mas- the downtown area, the distribution covers the entire sachusetts: U.S. Geological Survey Map GQ-463. study region. The SPT blow count data were analyzed CHUTE, N. E., 1966, Geology of the Norwood Quadrangle, Norfolk for susceptibility to liquefaction according to stan- and Suffolk Counties, Massachusetts: U.S. Geological Survey dard geotechnical procedures (Youd et al., 2001). Bulletin 1163, pp. B1–B78. Using the mapping criteria, the surficial geology CROSBY, I. B., 1923, The earthquake risk in Boston: Journal Boston Society of Civil Engineers, Vol. 10, No. 10, pp. 421–430. maps, and the geotechnical data, we prepared EBEL, J. E., 1996, The seventeenth century seismicity of liquefaction hazard maps for the greater Boston area. northeastern North America: Seismological Research Letters, These maps are appropriate for the design earthquake Vol. 67, No. 3, pp. 51–68. for Boston, MA (M6.0 and PGA 5 0.12g). Artificial EBEL, J. E., 1999, Speculations on the source parameters of the fill, marsh deposits, and beach deposits are mapped as 1638 earthquake in the northeastern using high hazard. Marsh deposits are loose deposits of silts inferences from modern seismicity: Seismological Research Letters, Vol. 70, No. 1, p. 114. and sands. The silty organic soils are not liquefiable, EBEL, J. E., 2000, A reanalysis of the 1727 earthquake at Newbury, and the sandy soils tend to be liquefiable during the Massachusetts: Seismological Research Letters, Vol. 71, design earthquake. Glaciofluvial deposits are mapped No. 3, pp. 364–374. as moderate hazard. Ground moraine and drumlin EBEL, J. E., 2006, The Cape Ann, Massachusetts, earthquake of th deposits are mapped as low hazard. 1755: A 250 anniversary perspective: Seismological Research Letters, Vol. 77, No. 1, pp. 74–86. HITCHCOCK,C.S.AND HELLEY, E. J., 2000, First Year Annual Report: Characterization of Subsurface Sediments for Lique- ACKNOWLEDGMENTS faction Hazard Assessment—Southern San Francisco Bay Area: National Earthquake Hazards Reduction Program, This research was supported by the U.S. Geological U.S. Geological Survey Award Number 99-HQ-GR-0097. Survey (USGS), Department of the Interior, under USGS HITCHCOCK, C. S.; LOYD, R. C.; AND HAYDON, W. D., 1999, award numbers 02HQGR0036 and 02HQGR0040. The Mapping liquefaction hazards in Simi Valley, Ventura county, California: Environmental Engineering Geoscience, views and conclusions contained in this document are Vol. 5, No. 4, pp. 441–458. those of the authors and should not be interpreted as JUANG, C. H.; JIANG, T.; AND ANDRUS, R. D., 2002, Assessing necessarily representing the official policies, either probability-based methods for liquefaction potential evalua-

Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16 15 Brankman and Baise

tion: Journal Geotechnical Geoenvironmental Engineering, TUTTLE,M.AND SEEBER, L., 1991, Historic and prehistoric Vol. 128, No. 7, pp. 580–589. earthquake-induced liquefaction in Newbury, Massachusetts: KAYE, C. A., 1961, Pleistocene Stratigraphy of Boston, Massachu- Geology (Boulder), Vol. 19, No. 6, pp. 594–597. setts: U.S. Geological Survey Professional Paper 424-B, TUTTLE, M.; SIMS, J. D.; AND ROY, D., 2000, Paleoseismology pp. B-73–B-76. Investigations in the Greater Boston, Massachusetts, Area: KAYE, C. A., 1978, Surficial Geologic Map of the Boston Area, Final Report for U.S. Geological Survey Award No. 1434- Massachusetts: U.S. Geological Survey Open-File Report 78- HQ-98-GR-00001. 111. TY, R. K. S., 1987, History and Characteristics of Man-Made Fill KAYE, C. A., 1982, Bedrock and Quaternary geology of the Boston in Boston and Cambridge: Unpublished M.S. Thesis, De- area, Massachusetts: Geological Society of America Reviews partment of Civil Engineering, Massachusetts Institute of in Engineering Geology, Vol. 5, pp. 25–40. Technology, Cambridge, MA. MONAHAN, P. A.; LEVSON, V. M.; MCQUARRIE, E. J.; BEAN, S. M.; U.S. GEOLOGICAL SURVEY (USGS), 2007, Conterminous States HENDERSON, P.; AND SY, A., 1998, Seismic microzonation Probabilistic Maps and Data: Electronic document, available mapping in Greater Victoria, British Columbia, Canada. In at http://earthquake.usgs.gov/research/hazmaps/products_data/ Dakoulas, P. and Yegian, M. (Editors), Geotechnical 48_States/index.php Earthquake Engineering and Soil Dynamics III: Geotechnical WHITMAN, R. V., 2002, Ground motions during the 1755 Cape Special Publication 75, American Society of Civil Engineers, Ann earthquake. In Earthquake Engineering Research In- Seattle, WA, pp. 128–140. stitute: Proceedings of the Seventh U.S. National Conference MONAHAN,P.A.;LEVSON,V.M.;HENDERSON,P.;AND SY, A., 2000, on Earthquake Engineering. Oakland, CA. Relative Liquefaction Hazard Map of Greater Victoria, British WOODHOUSE, D., 1989, The history of Boston: The impact of Columbia: Geoscience Map 2000-3a, British Columbia, Min- geology: Civil Engineering Practice, Vol. 4, No. 1, pp. 33–38. istry of Energy and Mines, Geologic Survey Branch, avail- WOODHOUSE, D.; BAROSH, P. J.; JOHNSON, E. G.; KAYE, C. A.; able at http://www.em.gov.bc.ca/DL/Hazards/liquefaction.pdf RUSSELL, H. A.; PITT, W. E., JR.; ALSUP, S. A.; AND FRANZ,K. OLDALE, R. N., 1962, Surficial Geology of the Reading Quadrangle, Massachusetts: U.S. Geological Survey Map GQ-168. E., 1991, Geology of Boston, Massachusetts, United States of America: Bulletin Association of Engineering Geologists, PLANT, M., 1742, A journal of the shocks of earthquakes felt near Newbury in New-England, from the year 1727 to the year Vol. 28, No. 4, pp. 375–512. 1741. Communicated in a letter from the Reverend Mr. YOUD,T.L.AND PERKINS, D. M., 1978, Mapping liquefaction- Matthias Plant to the Reverend Dr. Bearcroft: Philosophical induced ground failure potential: Journal Soil Mechanics Transactions of the Royal Society of London (1663–1775), Foundations Division, Vol. 104, No. 4, pp. 433–446. Vol. 42 (1742–1743), pp. 33–42. YOUD,T.L.AND PERKINS, D. M., 1987, Mapping of liquefaction SEASHOLES, N., 2003, Gaining Ground: MIT Press, Cambridge, MA. severity index: Journal Geotechnical Engineering, Vol. 113, SEED,H.B.AND IDRISS, I. M., 1971, Simplified procedure for No. 11, pp. 1374–1392. evaluating soil liquefaction potential: ASCE Journal of Soil YOUD, T. L.; IDRISS, I. M.; ANDRUS, R. D.; ARANGO, I.; CASTRO, Mechanics, Vol. 97(SM9), pp. 1249–1273. G.; CHRISTIAN, J. T.; DOBRY, R.; FINN, W. D. L.; HARDER,L. THOMPSON, W. B.; CHAPMAN, W. F.; BLACK, R. F.; RICHMOND,G. F., JR.; HYNES, M. E.; ISHIHARA, K.; KOESTER, J. P.; LIAO,S. M.; GRANT, D. R.; AND FULLERTON, D. S., 1991, Quaternary S. C.; MARCUSON, W. F., III; MARTIN, G. R.; MITCHELL, J. K.; geologic map of the Boston 4 3 6 quadrangle, United States MORIWAKI, Y.; POWER, M. S.; ROBERTSON, P. K.; SEED, R. B.; and Canada. In Quaternary Geologic Atlas of the United AND STOKOE, K. H. I. I., 2001, Liquefaction resistance of soils: States: U.S. Geological Survey Map I-1420 (NK-19). Summary report from the 1996 NCEER and 1998 NCEER/ THORSON, R. M., 2001, Remote aquifer response to the 18 NSF Workshops on Evaluation of Liquefaction Resistance November 1755 Cape Anne earthquake: Seismological Re- of Soils: Journal Geotechnical Geoenvironmental Engineering, search Letters, Vol. 72, No. 3, pp. 401–403. Vol. 127, No. 10, pp. 817–833.

16 Environmental & Engineering Geoscience, Vol. XIV, No. 1, February 2008, pp. 1–16